Dose intensification in acute myloid leukaemia: greater effectiveness at lower cost. Principal report of the Medical Research Council's ATML9 study

Between 1984 and 1990, 972 patients aged 1-79 years with acute myeloid leukaemia (AML), from 85 British hospitals, were entered into the MRC's 9th AML trial. Patients were randomized between DAT 1 + 5 (daunorubicin for 1 d, with cytarabine and 6-thioguanine for 5 d) and DAT 3 + 10 (same dose drugs for 3 and 10 d respectively) as induction therapy. The 63% who achieved complete remission (CR) were randomized to receive two courses of DAT 2 + 7 alternating with two courses of either MAZE (m-AMSA, 5-azacytidine, etoposide) or COAP (cyclophosphamide, vincristine, cytarabine, prednisone). Finally, those still in CR were randomized to receive either 1 year of maintenance treatment with eight courses of cytarabine and thioguanine followed by four courses of COAP, or no further cytotoxic therapy. Resistance to induction therapy was less common with the DAT 3 + 10 regimen than with DAT 1 + 5 (13% v 23%; P = 0.0001) and hence, despite a 5% increase in the risk of induction death, the CR rate was higher (66% v 61%; P = 0.15). Moreover, CR was achieved more rapidly with DAT 3 + 10 (median 34 v 46 d; P < 0.0001) and thus patients required less time in hospital (mean 20 v 29 d) and less blood product support. 5-year relapse-free survival (28% v 23%; P = 0.05) and survival (23% v 18%; P < 0.05) were also better with DAT 3 + 10. Post-remission intensification of therapy with MAZE resulted in fewer relapses (66% v 74% at 5 years; P = 0.03) but patients allocated MAZE required considerably more supportive care and 14 (4.5%) died following 312 MAZE courses, whereas no deaths occurred following COAP. 5-year survival was not significantly higher with MAZE (37% v 31%). Finally, although 1 year of outpatient maintenance treatment appeared to delay, but not prevent, recurrence it did not improve 5-year survival which was non-significantly worse for those allocated maintenance treatment (41% v 44%). We conclude that the more intensive induction regimen, DAT 3 + 10, is not only more effective than DAT 1 + 5, even for older patients, but is also less expensive; intensive post-remission therapy with MAZE achieves better leukaemic control but at the cost of substantial toxicity; whereas low-level maintenance therapy confers no apparent advantage in survival as well as being inconvenient and costly.